Infrared Imaging Video Bolometer with a Double Layer Absorbing Foil

Plasma and Fusion Research: Regular Articles Volume 2, S1052 (2007)

Igor V. MIROSHNIKOV, Artem Y. KOSTRYUKOV and Byron J. PETERSON1) St. Petersburg State Technical University, 29 Politechnicheskaya Str., St. Petersburg, 195251, Russia. 1)National Institute for Fusion Science, 322-6 Oroshi-cho, Toki, 509-5292, Japan (Received 30 November 2006 / Accepted 11 August 2007)

The object of the present paper is an infrared video bolometer with a bolometer foil consisting of two layers: the first layer is constructed of radiation absorbing blocks and the second layer is a thermal isolating base. The absorbing blocks made of a material with a high photon attenuation coefficient (gold) were spatially separated from each other while the base should be made of a material having high tensile strength and low thermal conductance (stainless steel). Such a foil has been manufactured in St. Petersburg and calibratedinNIFSusinga vacuum test chamber and a laser beam as an incident power source. A finite element method (FEM) code was applied to simulate the thermal response of the foil. Simulation results are in good agreement with the experimental calibration data. The temperature response of the double layer foil is a factor of two higher than that of a single foil IR video bolometer using the same absorber material and thickness. c 2007 The Japan Society of Plasma Science and Nuclear Fusion Research Keywords: plasma bolometry, infrared imaging bolometer, double layer foil, finite element method simulation DOI: 10.1585/pfr.2.S1052 1. Introduction 2. Double Layer Foil Design and Man- The idea of infrared imaging bolometry is to absorb ufacturing the incident plasma radiation in an ultra thin (1 µm-2.5 µm) The idea of DLF design is shown in Fig. 1. metal foil, to measure the temperature rise of the foil re- The first layer contains radiation absorbing blocks motely by means of infra red video camera and finally to made of a materialwith a high photon attenuation coeffi- calculate the incident radiation power flux as a function cient (gold, platinum). The blocks were spatially separated of the measured temperature rise. Initially [1] metal foil from each other while the base can be made of a material pixels were thermally isolated by two segmented support- having high tensile strength and low thermal conductance ing plates. The central part of the plates was later elimi- (stainless steel, havar). nated [2] and a large (10 cm×10 cm) foil was supported by Since the stainless steel SS304 thermal conductivity is acopper frame. The thermal fluxes between the neighbor- 30 times lower than that of gold evena1mmwidestain- ing pixels were calculated and taken into account during a less steel gaps between 5 mm × 5mmsquare golden blocks special reconstruction procedure. make it possible to decrease the effective thermal conduc- To achieve the most effective transformation of the tivity of the foil substantially. incident radiation power flux into infrared radiation mea- On the other hand the additional heat capacitance of sured by an infrared camera one should provide good ther- the base layer inevitably delays the thermal response of the mal isolation of each bolometer pixel from the other pixels DLFtothe incident radiation. The base should be therefore and from the supporting frame. The object of the present made as thin as possible to reduce that delay. paper is a first experimental sample of an infrared video Unfortunately,theheat capacitances (joule/gK) of ap- bolometer with a bolometer foil consisting of two layers: propriate base layer materials are higher than that of gold radiation absorbing layer and thermal isolating base. Part 2 describes the double layer foil (DLF) design and manufacturing procedures. Part 3 explains the finite element method used for thermal simulation of the DLF. Part 4 describes the experimental calibration test bench. Part 5 contains a comparison of the FEM simulation and experimental results and discussion. Conclusions and fu- ture plans are given in Part 6.

author’s e-mail: [email protected] Fig. 1 Double layer foil structure.

S1052-1 c 2007 The Japan Society of Plasma Science and Nuclear Fusion Research Plasma and Fusion Research: Regular Articles Volume 2, S1052 (2007)

Fig. 2 Cross section of a double layer foil with perforated base. and platinum. Fortunately, those materials are mechani- cally much stronger than noble metals, so that the base layer thickness can be reduced to microns at a large foil size. The first prototype used a 100 mm × 200 mm×8mi- crons SS304 foil. Then the foil was etched down to 2.5 microns using specially developedelectrochemical etching technology. The foil thickness has been measured by a me- chanical micrometer with an accuracy of ±0.2 micron. At this thickness the foil remained strong enough and could Fig. 3 DLF experimental sample. Front (radiation absorbing) be handled easily. Further improvement of the DLF im- side. plies SS foil etching down to 1micronwhile the main goal is to avoid etching non-uniformity. An alternative way to reduce the base layer heat capacitance is to etch away the parts of the base layer which lie beneath the absorbing blocks (see Fig. 2). Effective electrochemical etching cannot be used for that since it requires the foil to be put onto a special support and submerged into the electrolyte solution. Therefore the foil was installed in a stainless steel frame similar to those used in the IRVB experiments on LHD and the process was performed by a 0.6 keV Argon ion beam patterned by aspecial mask. The ion beam etching process is slow and needs to be improved. Gold deposition was done by means of a vacuum evaporation process with the use of a perforated molyb- denum mask put in front of the foil. The process makes it possible to vary the thickness of gold across the foil by putting additional covers onto the mask. Gold adhesion to Fig. 4 DLF experimental sample. Back (IR camera viewed) stainless steel is good enough. However careful ion beam side. cleaning of the surface is needed to avoid gold exfoliation. Thermal strain at the gold/SS interface can be totally re- lieved by ion beam bombardment as well. 3. Finite Element Method Thermal Fig. 3 shows the front (radiation absorbing) side of Simulation of the DLF the DLF prototype. Two golden blocks in the lower left The temperature distribution as a function of the inci- corner exfoliated. The blocksinlefthalf ofthefoilare dent radiation power can be calculated for a double layer 2.1 ± 0.1 microns thick, the gold thickness in the right part foil using a finite element method calculation. is 0.85 ± 0.1 micron. Four blocks in each part of the foil To ensure good accuracy the element size L should be were blackened by carbon spray (0.2-0.4 microns). chosen to be much smaller that the absorbing block size. Fig. 4 shows the back (IR camera) side of the DLF The time step ∆tshould be much smaller than the response prototype. The stainless steel base foil underneath the gold time of the foil to avoid calculation instability. absorber blocks was partially etched through the mask. A The temperature rise of a particular (x, )element can portion of the etching productswasdeposited back to the be calculated asfollows: foil under the mask (see the dark pattern). Four blocks in here each part of the foil were blackened by carbon spray.

2λx,λx±1, 2λx,λx,±1 Fx, + R(Tt,x, − Tr) + (Tt,x±1, − Tt,x,) + (Tt,x,±1 − Tt,x,±1) T + , , − T , , λ , + λ ± , λ , + λ ,± t 1 x t x = x x 1 x x 1 ∆t Cx,

S1052-2 Plasma and Fusion Research: Regular Articles Volume 2, S1052 (2007)

2 Fx, = PL -incident energy flux, data were processed to obtain temperature rise curves re- P -incident radiation power flux, sponding to the laser beam shutter opening. = σ ε + ε 3 R S ( ) f b)4Tr -energy losses due to infrared radiation 5. Results and Discussion σ = 5.67 10−8 W/m2K4, ε = 1-emissivity of the black- b Fig. 5 shows the comparison of the DLF and golden ened side of the foil and ε -emmisivity of the front side. f foil temperature responses to the laser shutter opening. The Since T − T T a linear approximation can be used: r r local temperature at the point where the laser beam was = σ ε + ε 4 − 4 pointed has been measured and calculated. Qr S ( f b)(T Tr ) Experimental and simulated spatial temperature pro- ≈ σS (ε + ε )4T 3(T − T )) f b r r files across the golden foil are shown in Fig. 6. The same 2 2 Cx, = CabsρabsL ha+Cwa f ρwa f L hwa f -total heat capaci- results obtained for the DLF are shown in Fig. 7. tance of the foil element One can see a reasonable agreement between the ex- ff Cabs, ρabs, ha absorbing block heat capacitance coefficient, perimental and simulation results. The di erence may be mass density and thickness. caused by the relatively large size of the finite element × Cwa f , ρwa f , hwa f -base layer heat capacitance coefficient, (0.5 mm 0.5 mm) and possible errors in the foil emissiv- mass density and thickness. ity measurements. λx, = (λabshabs + λwa f hwa f )-effective thermal conductance The gaps between the golden absorbing blocks on the of the foil element back side of the DLF were not blackened. That is why λabs, λwa f -absorbing block and base layer heat conduc- some “low temperature” zones have been observed (see tancecoefficients. blue line inFig.7). It is assumed that the temperature is 300 K at the moment Typical IRVB experimental conditions differ from the in time when the radiation flux is switched on. calibration procedure described above. The incident radi- Tx, = 300 K for t = 0 The numerical data was taken from the free internet data source [3].

4. Calibration Test Bench Both the double layer foil and the 2.5 microns golden IRVB foil were calibrated inavacuum test chamber by a HeNe laser. The vacuum chamber with the foil installed was evac- uated down to < 1mTorr. BoththeDLFand the golden foil thermal response increased by a factor of 2 when the pressure dropped from 100 to 1 mTorr. The effect is likely due to water evaporation from the blackened surface of the foil. Fig. 5 Thermal response of the DLF and 2.5 microns golden foil The fast framing 320 × 256 pixels Indigo-Phoenix IR to the 18 mW laser beam shutter opening. FEM simula- camera by FLIR was calibrated using a thermocouple and tion and experimental data. installed 1.1 m fromthebolometer foil in front of a ZnSe vacuum window. The camera spectral range is 3-5 microns and the temperature resolution is less than 0.5 K. The laser beam was pointed on the opposite side of the foil. The measured laser beam total power was 18 mW while the beam diameter was approximately 2 mm. Laser beam power and beam profile was measured by standard power meter and beam profiler. Aspecial experiment was done to measure the emissivity of the blackened surface. The metal sample surface was blackened and a black hole wasmade in the sample to provide an absolute black body radiation (emissivity = 1). The surface emissivity was found to be 0.87-0.88. Two types of experimental data have been recorded: asingle frame IR image and a sequence of images with a Fig. 6 Time evolution ofthegolden foil temperature profiles. frame rate of 400-450 Hz. Single images were used to see FEM simulation and experimental data. the saturated temperature profile across the foil. Sequence S1052-3 Plasma and Fusion Research: Regular Articles Volume 2, S1052 (2007) ation is not focused in a single point but spread across the Fig. 8 shows the comparison of the temperature rise whole foil. An additional experiment was done to see the curves measured experimentally. response of the golden foil and the DLF to the wide spread It is seen that the DLF is less sensitive for the signals laser beam (laser beam diameter −5 mm). faster than 500 ms and more sensitive for slow signals. The DLF sensitivity can be further improved by making the base foil thinner (current thickness is 2.5 micron) and/or by etching away the base layer underneath the absorbing blocks. For ITER applications where the thickness of the gold absorbers should be increased to at least 10 microns [4, 5] the DLF advantage is more sufficient. Let us compare the FEM simulated response of a 10 µm gold/ 2 µmstain- less steel DLF and 10 µm golden foil applicable to ITER (see Fig. 9). The fast thermal response of the golden foil and DLF are of the same order of magnitude while the slow thermal response of the DLF is 10 times higher than that of the golden foil.

6. Conclusions The double layer foil for an infrared video bolometer is a simply produced material which allows us to improve Fig. 7 Time evolution of the DLF temperature profiles. FEM the IRVB sensitivity. A first prototype of the DLF has simulation and experimental data. shown thermal response equal to that of similar golden foil at 500 ms signal rise time. The DLF sensitivity for slow signals is twice better than golden foil sensitivity. Further improvement should be achieved by use of a thinner (1 micron) base foil. Asimplerectangular grid finite element method simulation of the foil response was shown to be an adequate instrument to predict foil thermal properties. DLF in its presentembodiment does not allow a change of the size of the bolometer pixel (part of the foil considered as one bolometer “channel”). The problem can be solved by miniaturizing the DLF pattern. Another DLF drawback is the difference in gold and stainless steel at- Fig. 8 Thermal response of the DLF and 2.5 microns golden foil tenuation of high energy photons and particles which may to the 14.7 mW laser beam shutter opening, experimental provide slight spectral variations of the bolometer sensitiv- data. ity.

Fig. 9 Thermal response of the 10 µm gold/2 µmstainless steel DLF and 10 µm golden foil to the 14.7 mW/5mmlaser beam shutter opening. FEM simulation.

S1052-4 Plasma and Fusion Research: Regular Articles Volume 2, S1052 (2007)

Acknowledgements [3] www.matweb.com This research is supported by MEXT Grant-in-Aid [4] G.A. Wurden, B.J. Peterson, Imaging Bolometer Develop- # 16082207. ment for Large Fusion Devices, 1997 Varenna ITER Diag- nostic Workshop, LA-UR-97-3586. [5] G. Wurden, A rad-hard, steady-state, digital imaging [1]G.A.Wurden, B.J. Peterson and S. Sudo, Rev. Sci. Instrum. bolometer system for ITER, In Diagnostics for Experimen- 68, 766 (1997). talThermonuclear Fusion Reactors (Plenum Press, NY, [2] B.J Peterson, A. Yu. Kostryukov et al.,Rev.Sci. Instrum. 1996) pp.603-606. 74, 2040 (2003).

S1052-5